Molecular mRNA markers to detect ... peripheral blood from urothelial carcinoma patients

153
Molecular mRNA markers to detect and monitor circulating tumour cells in
peripheral blood from urothelial carcinoma patients
Mercedes Marín-Aguilera
a,b
, Lourdes Mengual
a,b
, Moisès Burset
a,b
, José Ríos
c
, Cristina
Gázquez
a
, Elisabet Ars
b
, María José Ribal
a
, Humberto Villavicencio
b
and Antonio
Alcaraz
a*
a
Laboratory and Department of Urology. Hospital Clínic, Institut d’Investigacions
Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain.
b
Laboratory of Molecular Biology, Department of Urology. Fundació Puigvert,
Barcelona, Spain.
c
Statistics and Methodology Support Unit. Hospital Clínic, Barcelona, Spain.
*Corresponding author:
Dr. Antonio Alcaraz
Hospital Clínic
Urology Department
Villarroel, 170
08036 Barcelona
Tel: [phone]
Fax: [phone]
E-mail: [e-mail]
Keywords: Carcinoma; Transitional Cell; Metastasis; cDNA pre-amplification; Real
Time PCR; Tumour Markers
Take-home message: Gene expression analysis of six specific bladder markers by the
high-sensitive PA qRT-PCR method is not suitable to identify and monitor
haemotogenous BUC dissemination in cystectomy treated patients. PCR-based
technology should be abandoned as a candidate method for detecting such tumour
spread.

154
ABSTRACT
Background: Detecting haematogenous spread of bladder urothelial carcinoma (BUC)
still represents a major challenge in clinical practice.
Objective: To test the efficiency of six mRNA bladder markers in detecting and
monitoring BUC haematogenous dissemination.
Design, Setting, and Participants: Between 2002 and 2004, 317 peripheral blood
(PB) samples were collected from 101 patients with BUC that underwent radical
cystectomy and from 29 healthy controls. Sequential PB sampling was performed in
patients at time of surgery and at 6, 12, 18 and 24 months afterwards. Median patients’
follow-up was 35 months.
Measurements: The presence of disseminated urothelial cells in PB was evaluated by
measuring gene expression from pre-amplified (PA) cDNA of KRT20, FXYD3,
C10orf116, UPK2, AGR2 and KRT19 genes, using quantitative RT-PCR (qRT-PCR).
Gene expression data were correlated with patients’ follow-up and outcome.
Results and limitations: The PA qRT-PCR method with selected markers in PB was
able to detect 2 cells in between 10
6
-10
7
leucocytes, depending on the gene. Differential
gene expression was found between patients and controls for FXYD3 (p=0.002),
C10orf116 (p=0.001), AGR2 (p=0.003) and KRT19 (p<0.001) genes. In contrast, no
differences were found between gene expression from metastasized and recurrence-
free patients (p=ns). Molecular monitoring of clinically metastasized patients did not
show an increased gene expression for selected markers in blood throughout the
follow-up after surgery.
Conclusions: Gene expression analysis of PB samples using six specific bladder
mRNA markers can distinguish between BUC patients and controls. However, despite
its high sensitivity, PA qRT-PCR with these potential markers is not a suitable
approach for detecting and monitoring BUC haematogenous dissemination. Thus,
PCR-based technology should be abandoned as a candidate method for detecting
such tumour spread.

155
1. Introduction
Bladder urothelial carcinoma (BUC) is one of the most aggressive epithelial tumours,
with a high rate of haematogenous and lymphatic dissemination. Cystectomy is the
gold standard of treatment for patients with clinically localized muscle-invasive BUC.
However, despite this radical treatment, around 50% of patients die within 5 years of
follow-up [1]. Such a high treatment failure rate is usually due to occult metastatic
disease which is not detected by conventional staging at the time of treatment [2].
Although lymph-node metastasis often represents an early prognostic indicator of
tumour invasiveness and metastatic dissemination, several works have focused on
detecting disseminated BUC cells in peripheral blood (PB) to monitor the metastatic
progression and response to chemotherapy [3-6]. Detecting such tumour dissemination
could be used in planning the optimal treatment and follow-up after radical cystectomy.
However, none of these strategies for detecting tumour cells in PB have yet been
incorporated into clinical practice.
Recently, we have described a list of bladder specific mRNA markers obtained from the
comparison of microarray data from bladder and PB tissues. Selected markers from
this analysis (KRT20, FXYD3, C10orf116, UPK2, AGR2 and KRT19) were used to
detect lymphatic BUC dissemination by quantitative RT-PCR (qRT-PCR), showing a
high sensitivity and specificity detecting disseminated tumour cells [7]. QRT-PCR
provides the advantages of being quantitative and more sensitive than conventional
RT-PCR in detecting rare events [8]. In addition, the sensitivity of qRT-PCR can further
be increased by a novel cDNA pre-amplification (PA) method that allows increasing the
quantity of specific cDNA targets prior to gene expression analysis [9;10].
In the present work we test the usefulness of six highly specific bladder markers in
detecting and monitoring the development of BUC dissemination in PB by a method
that combines PA of cDNA and qRT-PCR. The sensitivity of this method was assessed
by PB spiking experiments with dissociated cells from a BUC specimen. The clinical

156
applicability of this highly sensitive test was determined correlating gene expression
values with patients’ follow-up and outcome.
2. Material and Methods
2.1 Patients and samples
A total of 317 PB samples from 101 patients (10 women and 91 men; average 69 yr;
range 44-86 yr) who underwent radical cystectomy and pelvic lymphadenectomy
between August 2002 and July 2004 were included in this study. Only patients with
transitional cell carcinoma were considered. Five and 30 patients underwent
neoadjuvant and adjuvant chemotherapy, respectively. Six patients underwent
chemotherapy as a treatment for the metastasis. Clinical and pathological stages of
patients’ tumours [11;12] are shown in table 1. PB samples (8-10 mL) were collected
into EDTA-containing blood collection tubes (Sarstedt Monovette, Nümbrecht-
Rommelsdorf, Germany) just before the cystectomy (101 samples) and subsequently
at 6 (73 samples), 12 (55 samples), 18 (49 samples) and 24 (39 samples) months after
surgery. As controls, PB from 29 healthy volunteers (23 women and 6 men; average
age 32; range 22-62 yr) were also collected. To avoid epithelial contamination by the
skin during venipuncture, the first 5 mL of blood was discarded. The hospital ethics
committee approved this study and the patients and controls provided their informed
consent before participating in the study.
2.2 Marker selection
To determine a panel of highly specific mRNA markers for BUC dissemination, gene
expression from bladder tissue (normal and tumour) and blood were compared by
microarrays analysis [7]. Those four genes with the highest expression in urothelial
tissue but lowest expression in blood (FXYD3, C10orf116, AGR2, and KRT19) were
selected as potential candidates for haematogenous BUC dissemination. KRT20 and
UPK2, two conventional markers for epithelial cells that have been widely tested as
BUC dissemination markers in the bibliography, were also selected for study [13-16].
2.3 RNA extraction and Reverse Transcription

157
PB specimens were layered onto 10 mL of Ficoll-Paque (GE Healthcare Life Sciences,
Uppsala, Sweden). Mononuclear cells were isolated and total RNA was extracted using
the Trizol Reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) according to the
manufacturer’s instructions. Quantity of total RNA was assessed by absorbance at 260
nm in an Ultrospec 3100 pro spectrophotometer (GE Healthcare Life Sciences,
Uppsala, Sweden). Five µg of total RNA were reverse-transcribed with a random
hexamer primer mix in a 20 µl reaction mix using SuperScriptII Reverse Transcriptase
(Invitrogen Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s
instructions.
2.4 Pre-amplification and qRT-PCR analysis
cDNA from control PB specimens (n=29), from PB samples obtained at time of the
surgery (n=101) and from samples obtained at 6, 12, 18 and/or 24 months after the
surgery from patients who recurred or died because of the cancer (n=47) were pre-
amplified. This pre-amplification (PA) was performed by means of a multiplex PCR with
the 6 specific bladder mRNA markers (KRT20, FXYD3, C10orf116, UPK2, AGR2 and
KRT19) and the endogenous control GUSB, using the TaqMan PreAmp Master Mix Kit
following manufacturer’s instructions (Applied Biosystems), except that the final volume
of the reaction was 25 µl. Briefly, the 7 TaqMan Gene Expression Assays (Applied
Biosystems) were pooled together at 0.2X of final concentration. Subsequently, 6.25 µl
of this assay mix were combined with 1.25 µl of cDNA (equivalent to 125 ng) and 12.5
µl of the TaqMan PreAmp Master Mix (2X) in a final volume of 25 µl. Thermal cycling
conditions were as follows: initial hold at 95ºC during 10 min and ten pre-amplification
cycles of 15 sec at 95ºC and 4 min at 60ºC. Gene expression quantification of the
selected genes and endogenous control was performed using an ABI PRISM 7000
Sequence Detection System (Applied Biosystems, Foster City, CA, USA) according to
the manufacturer’s recommendations, except that the final volume of the reaction was
20 µL. Commercially available primers and Minor Groove Binder (MGB) probes for all